Regex in dataframe (Pandas)

Question

I have a dataframe with a column like that

COL1
PACK[30% /2 prod.(if fidelity)]
PACK[3.85 € /2 prod.(if fidelity)]
PACK[40% /2nd prod.]
PACK[3.5 € /2 prod.]

I want create other column as following according COL1

fidelity_perc fidelity_euro rem_perc  rem_eu
30           3,85           40        3,5

using regex.

For PACK[40% /2nd prod.] I did (?<=PACK\[)\d+(?=%) but it's also walk for PACK[30% /2 prod.(if fidelity)] and I don't want this.

Maybe you want df['COL1'].str.extract(r'PACK\[(\d[\d.]*)%[^][()]*]', expand=False)? See regex101.com/r/3jROEh/2 — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Dec 2, 2020 at 14:44
@Wiktor Thank you, and I want to add contains "fidelity" in my regex for example how could I do ? — T.mum
– T.mum, Commented Dec 2, 2020 at 14:51
df['fidelity_perc'] = df['COL1'].str.extract(r'PACK\[(\d[\d.]*)%[^][()]*]', expand=False)? — Wiktor Stribiżew
– Wiktor Stribiżew, Commented Dec 2, 2020 at 14:53
@Wiktor This is ok for PACK[40% /2nd prod.] but not for PACK[30% /2 prod.(if fidelity)], I have to speficy contains "fidelity" — T.mum
– T.mum, Commented Dec 2, 2020 at 14:55

Wiktor Stribiżew · Accepted Answer · 2020-12-02 15:58:54Z

2

To get the PACK[40% /2nd prod.], you can use

df['COL1'].str.extract(r'PACK\[(\d[\d.]*)%[^][()]*]', expand=False)

To get the PACK[30% /2 prod.(if fidelity)], you can use

df['COL1'].str.extract(r'PACK\[(\d[\d.]*)%[^][]*fidelity[^][]*]', expand=False)

Similarly, if you need to only find a value if there is a + sign, just replace fidelity with \+:

df['COL1'].str.extract(r'PACK\[(\d[\d.]*)%[^][]*\+[^][]*]', expand=False)

Details

PACK\[ - a literal PACK[ string
(\d[\d.]*) - Capturing group 1: a digit an then any zero or more digits or dots as many as possible
% - a % sign
[^][]* - any zero or more chars other than [ and ] as many as possible
\+ - a plus
fidelity - a fidelity substring
[^][]* - any zero or more chars other than [ and ] as many as possible
[^][()]* - any zero or more chars other than [, ], ( and ) as many as possible
] - a ] char.

answered Dec 2, 2020 at 15:58

Wiktor Stribiżew

631k41 gold badges502 silver badges633 bronze badges

Sign up to request clarification or add additional context in comments.

4 Comments

Hello I have an other question ! if want to add "not contains +" how could I do?

@T.mum You add + into the negated character class.

what's the negated class?

@T.mum See here.